U.S.-E.U. Coordination on Artificial Intelligence Shouldn’t Ignore U.S. Voters’ Concerns About Jobs

United States and European Union find common ground on AI regulation

Artificial intelligence is a relatively new focus of bilateral tech discussions between the United States and the European Union. In contrast to issues like data transfers, which have been evergreen (and ever problematic) since the late ’90s, attempts to establish joint norms on AI usage are fairly recent, only gaining steam in the last couple of years. For its part, the European Union has already staked out a clear position on AI regulation via its draft人工智能法: a risk-based method whereby regulators group AI systems based on the level of risk they pose and regulate them accordingly (or require companies to self-regulate).

Coming out of September 2021’s inaugural meeting of the U.S.-E.U. Trade and Technology Council, a trans-Atlantic forum that aims to facilitate cooperation across a variety of issues, the United States and European Union resolved to “develop and implement AI systems that are innovative and trustworthy and that respect universal human rights and shared democratic values.”

这导致了一个星期的后续会议nouncement of a “joint road map on evaluation and measurement tools for trustworthy AI and risk management, as well as a common project on privacy-enhancing technologies.” These announcements point to increasing trans-Atlantic4. This may be a net positive for companies deploying AI applications on both sides of the Atlantic, as they would benefit from policy harmonization, but it could also entail new restrictions for U.S. companies using AI at home if the United States follows the European Union’s tougher line on AI regulation.

U.S. voters are worried about privacy and AI bias, but they’re more worried about jobs

但事实证明美国选民不太关注about whether AI will violate their democratic values — like the right to privacy, as well as racial and gender equality — than they are about whether it will take their jobs. Two-thirds of U.S. voters believe that AI could eliminate jobs and cause unemployment. A smaller but still sizable share of voters (54%) agree it could lead to privacy violations. The shares who agree that AI could lead to biases against women or people from certain racial and ethnic backgrounds fall much further down the list. In the context of ongoing discussions on AI regulation under the auspices of the TTC, the gap suggests that U.S. voters would prefer the Biden administration to focus first on ensuring that AI does not lead to unemployment and address challenges related to privacy and certain types of bias second.

U.S. voters still favor cooperating with the European Union on AI standards

The share of voters concerned about job loss and unemployment also outweighs the more than half of voters who see positive benefits arising from AI, like the potential to create jobs in new industries or to solve intractable problems in mathematics and health care. This gap likely informs voters’ strong preference that the United States and European Union establish principles for the ethical use of AI rather than allow companies on both sides of the Atlantic to make and sell AI applications with no ex-anteregard for ethical concerns.

U.S. voters’ interest in seeing trans-Atlantic cooperation on AI ethics also leads the pack relative to a host of other values-based issues under discussion within the TTC, even though only 29% of voters said they had heard “some” or “a lot” about artificial intelligence bias.

Companies face a critical juncture for shaping the United States’ approach to AI regulation

Tech or other companies deploying AI solutions in the United States should take note of its drift toward a risk-based approach in line with the AI Act that is making its way through the E.U. legislative process. This development, and the work of the AI subgroup in the TTC, have a strong demand signal from U.S. voters.

At the same time, the tech industry is at a critical juncture for providing input that could, at the margins, shape the United States’ approach to AI regulation and its cooperation with Europe. As industry leaders engage with policymakers, they should deploy narratives about employment and privacy that will resonate with U.S. voters — a precondition for making regulatory progress in a midterm election year, amid a polarized U.S. political climate. This, in turn, could facilitate greater public-private dialogue on trans-Atlantic AI norms, both under the auspices of the TTC and elsewhere.